wehub-resource-sync
363 lines
13 KiB
Python
363 lines
13 KiB
Python
"""
|
|
切片API路由
|
|
"""
|
|
|
|
from typing import List, Optional
|
|
from fastapi import APIRouter, Depends, HTTPException, Query
|
|
from sqlalchemy.orm import Session
|
|
from ...core.database import get_db
|
|
from ...services.clip_service import ClipService
|
|
from ...schemas.clip import ClipCreate, ClipUpdate, ClipResponse, ClipListResponse, ClipStatus, ClipFilter
|
|
from ...schemas.base import PaginationParams
|
|
from ...models.clip import Clip
|
|
import logging
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
router = APIRouter()
|
|
|
|
|
|
def get_clip_service(db: Session = Depends(get_db)) -> ClipService:
|
|
"""Dependency to get clip service."""
|
|
return ClipService(db)
|
|
|
|
|
|
@router.patch("/{clip_id}/title", response_model=ClipResponse)
|
|
async def update_clip_title(
|
|
clip_id: str,
|
|
title_data: dict,
|
|
clip_service: ClipService = Depends(get_clip_service)
|
|
):
|
|
"""Update clip title."""
|
|
try:
|
|
new_title = title_data.get("title", "").strip()
|
|
if not new_title:
|
|
raise HTTPException(status_code=400, detail="标题不能为空")
|
|
|
|
if len(new_title) > 200:
|
|
raise HTTPException(status_code=400, detail="标题长度不能超过200个字符")
|
|
|
|
# 更新切片标题
|
|
clip = clip_service.update_clip(clip_id, ClipUpdate(title=new_title))
|
|
if not clip:
|
|
raise HTTPException(status_code=404, detail="切片不存在")
|
|
|
|
# 返回更新后的切片信息
|
|
return ClipResponse(
|
|
id=str(clip.id),
|
|
project_id=str(clip.project_id),
|
|
title=str(clip.title),
|
|
description=str(clip.description) if clip.description else None,
|
|
start_time=getattr(clip, 'start_time', 0),
|
|
end_time=getattr(clip, 'end_time', 0),
|
|
duration=int(getattr(clip, 'duration', 0)),
|
|
score=getattr(clip, 'score', None),
|
|
status=getattr(clip, 'status', 'pending'),
|
|
video_path=getattr(clip, 'video_path', None),
|
|
tags=getattr(clip, 'tags', []) or [],
|
|
clip_metadata=getattr(clip, 'clip_metadata', {}) or {},
|
|
created_at=getattr(clip, 'created_at', None),
|
|
updated_at=getattr(clip, 'updated_at', None),
|
|
collection_ids=[]
|
|
)
|
|
|
|
except HTTPException:
|
|
raise
|
|
except Exception as e:
|
|
logger.error(f"更新切片标题失败: {e}")
|
|
raise HTTPException(status_code=500, detail=f"更新切片标题失败: {str(e)}")
|
|
|
|
|
|
@router.post("/{clip_id}/generate-title", response_model=dict)
|
|
async def generate_clip_title(
|
|
clip_id: str,
|
|
clip_service: ClipService = Depends(get_clip_service)
|
|
):
|
|
"""Generate a new title for a clip using LLM."""
|
|
try:
|
|
# 获取切片信息
|
|
clip = clip_service.get(clip_id)
|
|
if not clip:
|
|
raise HTTPException(status_code=404, detail="切片不存在")
|
|
|
|
# 直接从clip_metadata获取内容，不需要从文件系统获取
|
|
clip_metadata = getattr(clip, 'clip_metadata', {}) or {}
|
|
|
|
if not clip_metadata:
|
|
raise HTTPException(status_code=404, detail="切片元数据不存在")
|
|
|
|
# 准备LLM输入数据
|
|
llm_input = [{
|
|
"id": clip_id,
|
|
"title": clip_metadata.get('outline', '') or getattr(clip, 'title', ''),
|
|
"content": clip_metadata.get('content', []),
|
|
"recommend_reason": clip_metadata.get('recommend_reason', '')
|
|
}]
|
|
|
|
# 调用LLM生成标题
|
|
from ...utils.llm_client import LLMClient
|
|
from ...core.shared_config import PROMPT_FILES
|
|
|
|
llm_client = LLMClient()
|
|
|
|
# 加载标题生成提示词
|
|
with open(PROMPT_FILES['title'], 'r', encoding='utf-8') as f:
|
|
title_prompt = f.read()
|
|
|
|
# 调用LLM
|
|
raw_response = llm_client.call_with_retry(title_prompt, llm_input)
|
|
|
|
if not raw_response:
|
|
raise HTTPException(status_code=500, detail="LLM调用失败")
|
|
|
|
# 解析LLM响应
|
|
titles_map = llm_client.parse_json_response(raw_response)
|
|
|
|
if not isinstance(titles_map, dict) or clip_id not in titles_map:
|
|
raise HTTPException(status_code=500, detail="LLM返回格式错误")
|
|
|
|
generated_title = titles_map[clip_id]
|
|
|
|
return {
|
|
"clip_id": clip_id,
|
|
"generated_title": generated_title,
|
|
"success": True
|
|
}
|
|
|
|
except HTTPException:
|
|
raise
|
|
except Exception as e:
|
|
logger.error(f"生成切片标题失败: {e}")
|
|
raise HTTPException(status_code=500, detail=f"生成切片标题失败: {str(e)}")
|
|
|
|
|
|
@router.post("/", response_model=ClipResponse)
|
|
async def create_clip(
|
|
clip_data: ClipCreate,
|
|
clip_service: ClipService = Depends(get_clip_service)
|
|
):
|
|
"""Create a new clip."""
|
|
try:
|
|
clip = clip_service.create_clip(clip_data)
|
|
# Convert to response schema
|
|
status_obj = getattr(clip, 'status', None)
|
|
status_value = status_obj.value if hasattr(status_obj, 'value') else 'pending'
|
|
|
|
return ClipResponse(
|
|
id=str(getattr(clip, 'id', '')),
|
|
project_id=str(getattr(clip, 'project_id', '')),
|
|
title=str(getattr(clip, 'title', '')),
|
|
description=str(getattr(clip, 'description', '')) if getattr(clip, 'description', None) else None,
|
|
start_time=getattr(clip, 'start_time', 0),
|
|
end_time=getattr(clip, 'end_time', 0),
|
|
duration=getattr(clip, 'duration', 0),
|
|
score=getattr(clip, 'score', None),
|
|
status=status_value,
|
|
video_path=getattr(clip, 'video_path', None),
|
|
tags=getattr(clip, 'tags', []) or [],
|
|
clip_metadata=getattr(clip, 'clip_metadata', {}) or {},
|
|
created_at=getattr(clip, 'created_at', None) if isinstance(getattr(clip, 'created_at', None), (type(None), __import__('datetime').datetime)) else None,
|
|
updated_at=getattr(clip, 'updated_at', None) if isinstance(getattr(clip, 'updated_at', None), (type(None), __import__('datetime').datetime)) else None,
|
|
collection_ids=[]
|
|
)
|
|
except Exception as e:
|
|
raise HTTPException(status_code=400, detail=str(e))
|
|
|
|
|
|
@router.get("/", response_model=ClipListResponse)
|
|
async def get_clips(
|
|
page: int = Query(1, ge=1, description="Page number"),
|
|
size: int = Query(20, ge=1, le=100, description="Page size"),
|
|
project_id: Optional[str] = Query(None, description="Filter by project ID"),
|
|
status: Optional[ClipStatus] = Query(None, description="Filter by status"),
|
|
clip_service: ClipService = Depends(get_clip_service)
|
|
):
|
|
"""Get paginated clips with optional filtering."""
|
|
try:
|
|
pagination = PaginationParams(page=page, size=size)
|
|
|
|
filters = None
|
|
if project_id or status:
|
|
filters = ClipFilter(
|
|
project_id=project_id,
|
|
status=status
|
|
)
|
|
|
|
return clip_service.get_clips_paginated(pagination, filters)
|
|
except Exception as e:
|
|
raise HTTPException(status_code=400, detail=str(e))
|
|
|
|
|
|
@router.get("/{clip_id}", response_model=ClipResponse)
|
|
async def get_clip(
|
|
clip_id: str,
|
|
clip_service: ClipService = Depends(get_clip_service)
|
|
):
|
|
"""Get a clip by ID."""
|
|
try:
|
|
clip = clip_service.get(clip_id)
|
|
if not clip:
|
|
raise HTTPException(status_code=404, detail="Clip not found")
|
|
return clip
|
|
except HTTPException:
|
|
raise
|
|
except Exception as e:
|
|
raise HTTPException(status_code=400, detail=str(e))
|
|
|
|
|
|
@router.put("/{clip_id}", response_model=ClipResponse)
|
|
async def update_clip(
|
|
clip_id: str,
|
|
clip_data: ClipUpdate,
|
|
clip_service: ClipService = Depends(get_clip_service)
|
|
):
|
|
"""Update a clip."""
|
|
try:
|
|
clip = clip_service.update_clip(clip_id, clip_data)
|
|
if not clip:
|
|
raise HTTPException(status_code=404, detail="Clip not found")
|
|
return clip
|
|
except HTTPException:
|
|
raise
|
|
except Exception as e:
|
|
raise HTTPException(status_code=400, detail=str(e))
|
|
|
|
|
|
@router.delete("/{clip_id}")
|
|
async def delete_clip(
|
|
clip_id: str,
|
|
clip_service: ClipService = Depends(get_clip_service)
|
|
):
|
|
"""Delete a clip."""
|
|
try:
|
|
success = clip_service.delete(clip_id)
|
|
if not success:
|
|
raise HTTPException(status_code=404, detail="Clip not found")
|
|
return {"message": "Clip deleted successfully"}
|
|
except HTTPException:
|
|
raise
|
|
except Exception as e:
|
|
raise HTTPException(status_code=400, detail=str(e))
|
|
|
|
|
|
@router.post("/cleanup-duplicates")
|
|
async def cleanup_duplicate_clips(
|
|
project_id: str,
|
|
db: Session = Depends(get_db)
|
|
):
|
|
"""清理项目中的重复切片数据"""
|
|
try:
|
|
from ...models.project import Project
|
|
import json
|
|
from pathlib import Path
|
|
from ...core.config import get_data_directory
|
|
|
|
# 获取项目
|
|
project = db.query(Project).filter(Project.id == project_id).first()
|
|
if not project:
|
|
raise HTTPException(status_code=404, detail="项目不存在")
|
|
|
|
# 获取数据库中的所有切片
|
|
db_clips = db.query(Clip).filter(Clip.project_id == project_id).all()
|
|
logger.info(f"数据库中有 {len(db_clips)} 个切片")
|
|
|
|
# 读取文件系统中的原始数据
|
|
data_dir = get_data_directory()
|
|
project_dir = data_dir / "projects" / project_id
|
|
clips_metadata_file = project_dir / "metadata" / "clips_metadata.json"
|
|
|
|
if not clips_metadata_file.exists():
|
|
raise HTTPException(status_code=404, detail="切片元数据文件不存在")
|
|
|
|
with open(clips_metadata_file, 'r', encoding='utf-8') as f:
|
|
original_clips = json.load(f)
|
|
|
|
logger.info(f"文件系统中有 {len(original_clips)} 个切片")
|
|
|
|
# 创建原始切片的ID映射
|
|
original_clip_ids = {clip['id']: clip for clip in original_clips}
|
|
|
|
# 清理重复数据
|
|
deleted_count = 0
|
|
kept_count = 0
|
|
|
|
for db_clip in db_clips:
|
|
metadata = db_clip.clip_metadata or {}
|
|
original_id = metadata.get('id')
|
|
|
|
if original_id and original_id in original_clip_ids:
|
|
# 这个切片是有效的，保留
|
|
kept_count += 1
|
|
logger.info(f"保留切片: {db_clip.title} (ID: {original_id})")
|
|
else:
|
|
# 这个切片是重复的或无效的，删除
|
|
logger.info(f"删除重复切片: {db_clip.title} (DB ID: {db_clip.id})")
|
|
db.delete(db_clip)
|
|
deleted_count += 1
|
|
|
|
db.commit()
|
|
|
|
return {
|
|
"project_id": project_id,
|
|
"project_name": project.name,
|
|
"original_count": len(original_clips),
|
|
"db_before_count": len(db_clips),
|
|
"kept_count": kept_count,
|
|
"deleted_count": deleted_count,
|
|
"message": f"清理完成：保留 {kept_count} 个，删除 {deleted_count} 个重复切片"
|
|
}
|
|
|
|
except HTTPException:
|
|
raise
|
|
except Exception as e:
|
|
logger.error(f"清理重复切片失败: {e}")
|
|
db.rollback()
|
|
raise HTTPException(status_code=500, detail=f"清理失败: {str(e)}")
|
|
|
|
|
|
@router.post("/resync-project")
|
|
async def resync_project_clips(
|
|
project_id: str,
|
|
db: Session = Depends(get_db)
|
|
):
|
|
"""重新同步项目的切片数据"""
|
|
try:
|
|
from ...models.project import Project
|
|
from ...services.data_sync_service import DataSyncService
|
|
from pathlib import Path
|
|
from ...core.config import get_data_directory
|
|
|
|
# 获取项目
|
|
project = db.query(Project).filter(Project.id == project_id).first()
|
|
if not project:
|
|
raise HTTPException(status_code=404, detail="项目不存在")
|
|
|
|
# 删除现有的切片数据
|
|
existing_clips = db.query(Clip).filter(Clip.project_id == project_id).all()
|
|
deleted_count = len(existing_clips)
|
|
for clip in existing_clips:
|
|
db.delete(clip)
|
|
db.commit()
|
|
logger.info(f"删除了 {deleted_count} 个现有切片")
|
|
|
|
# 重新同步数据
|
|
data_dir = get_data_directory()
|
|
project_dir = data_dir / "projects" / project_id
|
|
|
|
sync_service = DataSyncService(db)
|
|
synced_count = sync_service._sync_clips_from_filesystem(project_id, project_dir)
|
|
|
|
return {
|
|
"project_id": project_id,
|
|
"project_name": project.name,
|
|
"deleted_count": deleted_count,
|
|
"synced_count": synced_count,
|
|
"message": f"重新同步完成：删除 {deleted_count} 个，同步 {synced_count} 个切片"
|
|
}
|
|
|
|
except HTTPException:
|
|
raise
|
|
except Exception as e:
|
|
logger.error(f"重新同步切片失败: {e}")
|
|
db.rollback()
|
|
raise HTTPException(status_code=500, detail=f"重新同步失败: {str(e)}") |